Perception Engineer (Multimodal & Embodied AI)

Shield AI; California; On-site; USD 150,000 - 190,000; Full time

Benefits offered by this job

Excellent Medical Coverage

Mental Health Employee Assistance

Paid Parental Leave

Pet Insurance

Flexible Work Hours

Onsite Gym (DC)

Gym Discount (San Diego)

Free Parking

Competitive Compensation

Stock Benefits

401K Services and Match

Job summary

Shield AI’s Perception team seeks an engineer to translate multimodal vision advances into real-world autonomous aircraft and vehicles. You’ll prototype and evaluate ideas with researchers and autonomy engineers, addressing sensor, compute, and edge deployment challenges.

You will combine learned and classical techniques, optimize for edge use, and contribute Python/C++ software for production with experienced teammates.

Qualifications

• Experience building substantial research software.
• Research depth in multimodal vision, visual foundation models, or embodied learning.
• PhD in computer vision, ML, robotics, or equivalent record.
• Strong Python skills and PyTorch experience.
• Experience with robotics, real sensors, temporal data, or multi-sensor perception.
• Experience bringing a research prototype into sustained use.
• Experience with C++, edge inference, or model optimization (ONNX, TensorRT, CUDA).
• Familiarity with vision-language models or embodied AI.
• You don’t need to meet every qualification; we value depth and curiosity.

Responsibilities

• Evaluate promising approaches and build prototypes for real hardware.
• Turn advances in multimodal vision into practical capabilities for autonomous aircraft and vehicles.
• Collaborate with researchers and engineers to carry ideas into product (Hivemind).
• Combine learned and classical perception methods to tackle sensor challenges.
• Optimize and integrate capabilities for edge deployment.
• Contribute Python and C++ software with support from product engineers.
• Deliver tested, documented capabilities and support adoption by other teams.

Skills

Python

PyTorch

C++

Robotics perception

Edge inference

Model optimization

Vision-language models

Experimentation

Research software

Education

PhD in computer vision or ML

Tools

ONNX

TensorRT

CUDA

ROS

Job description

• Bring promising AI research into autonomous systems that operate in the real world. Shield AI's Perception team is looking for an engineer who enjoys exploring new ideas and making them work on real hardware. You'll help turn advances in multimodal computer vision, visual foundation models, and embodied AI into useful capabilities for autonomous aircraft and vehicles
• Working alongside researchers and autonomy engineers, you'll evaluate promising approaches, build prototypes, and carry successful ideas into the Hivemind product. You'll combine learned methods with established perception techniques, working through the sensor, software, and compute challenges that make edge robotics different from a benchmark
• Evaluate emerging models and methods against representative sensor data and autonomy needs
• Build prototypes and experiments that reveal strengths, limitations, and practical deployment opportunities
• Combine learned and classical approaches to improve detection, tracking, geometric vision, or sensor fusion
• Optimize and integrate capabilities for edge deployment, contributing Python and C++ software with support from experienced product engineers
• Help deliver tested, documented capabilities and stay involved as other teams adopt them

Benefits

• Excellent Medical Coverage
• Mental Health Employee Assistance Program
• Paid Parental Leave
• Pet Insurance
• Flexible Work Hours
• Onsite Gym (DC)
• Gym Discount (San Diego)
• Free Parking
• Competitive Compensation
• Stock Benefits
• 401K Services and Match

Experience building substantial research software, an interest in developing production-quality C++, and a collaborative approach to solving problemsResearch depth in an area such as multimodal vision, visual foundation models, video understanding, open-vocabulary perception, or embodied learningA PhD in computer vision, machine learning (ML/DL), robotics, or a related field, or an equivalent record of advanced research and hands-on engineeringA thoughtful approach to experiments, evaluation data, and understanding why models failStrong Python skills and experience with PyTorch or a comparable ML frameworkExperience with robotics, real sensors, temporal data, or multi-sensor perceptionExperience bringing a research prototype into sustained useExperience with C++, edge inference, or model optimization using tools such as ONNX, TensorRT, or CUDAFamiliarity with vision-language models, vision-language-action models, or embodied AIYou don't need to meet every preferred qualification or arrive as an expert C++ engineer. We welcome recent PhD graduates and early-career researchers, and we value technical depth, curiosity, and enthusiasm to transfer research to business value more than publication count alone